Commit History

  1. v2: S12 server-builder format, repairs, augmentation, template-level eval split, probe split (train_on/tags columns)

    Verified signature
    7bb0aa6

  1. docs: validation results of the LoRA trained on this revision (tool 20/20 x2, chat 20/20, 0 false calls, user name 40/40) (README)

    Verified signature
    f21709d

  1. feat: name augmentation — 12% rows (647/38) carry the server user-profile block and address the user by name (final_source +name) (README)

    Verified signature
    ed8d91d
  2. feat: name augmentation — 12% rows (647/38) carry the server user-profile block and address the user by name (final_source +name)

    Verified signature
    786abca
  3. feat: name augmentation — 12% rows (647/38) carry the server user-profile block and address the user by name (final_source +name)

    Verified signature
    a999728

  1. fix: 40 rows hand-corrected (false action claims, misread chit-chat, self-labeling, stutter) + 5 new validation rules (README)

    Verified signature
    b73266f
  2. fix: 40 rows hand-corrected (false action claims, misread chit-chat, self-labeling, stutter) + 5 new validation rules

    Verified signature
    bb2cfb8

  1. OpenMascotAI Jarvis v1 (2026-09-04): persona-free tool-calling dataset, 5,494/322 rows, assistant text regenerated by stock Qwen3.8-27B (README)

    Verified signature
    54624c9
  2. OpenMascotAI Jarvis v1 (2026-09-04): persona-free tool-calling dataset, 5,494/322 rows, assistant text regenerated by stock Qwen3.8-27B

    Verified signature
    7bb5d99
  3. OpenMascotAI Jarvis v1 (2026-09-04): persona-free tool-calling dataset, 5,494/322 rows, assistant text regenerated by stock Qwen3.8-27B

    Verified signature
    e33ef59
  4. initial commit

    Verified signature
    ba3111e